You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi, many thanks for your great work!
I'm trying to reproduce your result by training your algorithm with COIN dataset (/scripts/coin/live1+.sh)
Training has been finished for 50 hours by using just one single A100, and it seems that the evaluation script runs automatically once the training process is finished.
However, this evaluation process is too slow (about 20 sec/iter speed, and it takes more than 100 hours) when I use one single A100.
When we run live1+.sh script with multiple A100, evaluation processing time isn't more than 1 hour.
I'd really appreciate it if you could look into this issue.
The text was updated successfully, but these errors were encountered:
Yes, right.
When I run /scripts/coin/live1+.sh with only one single GPU, training time takes about 50 hours, but evaluation time takes too longer (about 100+ hours). Please see below image.
Hi, many thanks for your great work!
I'm trying to reproduce your result by training your algorithm with COIN dataset (/scripts/coin/live1+.sh)
Training has been finished for 50 hours by using just one single A100, and it seems that the evaluation script runs automatically once the training process is finished.
However, this evaluation process is too slow (about 20 sec/iter speed, and it takes more than 100 hours) when I use one single A100.
When we run live1+.sh script with multiple A100, evaluation processing time isn't more than 1 hour.
I'd really appreciate it if you could look into this issue.
The text was updated successfully, but these errors were encountered: